BR112016024543B1 - ACTIVE SEPARATOR AND SYSTEM WITH SERVICE CENTER - Google Patents

Abstract

SYSTEM AND APPARATUS FOR CALIBRATION OF OPTICAL MODULATION INDEX IN A CATV NETWORK. These are systems and methods for calibrating the optical modulation index in a CATV network.

Description

CROSS REFERENCE TO RELATED ORDERS

[001] This application claims the benefit of priority under 35 U.S.C. paragraph 119(e) to U.S. Provisional Application No. 62/043793 filed August 29, 2014, to U.S. Provisional Application No. 62/052213, filed September 18 2014, to the Provisional Application in U.S. 61/984303 filed on April 25, 2014, and to the Provisional Application in U.S. 61/982089, filed on April 21, 2014.

BACKGROUND

[002] The present disclosure relates to systems and methods that process signals over a cable transmission network.

[003] Although Cable Television (CATV) networks originally delivered content to subscribers over long distances using an RF-only transmission system, modern CATV transmission systems have replaced a large portion of the RF transmission path with a most effective optical network, which creates a hybrid transmission system in which cable content originates and terminates as RF signals over coaxial cables, but is converted to optical signals for transmission over the bulk of the intermediate distance between the content provider and the subscriber. Specifically, CATV networks include a head end at the content provider to receive RF signals representing many content channels. The service center receives the respective RF content signals, multiplexes them using an RF combining network, converts the combined RF signal into an optical signal (typically using the RF signal to modulate a laser ) and emits the optical signal to a fiber optic network that communicates the signal to one or more nodes, each of which brings together a group of subscribers. The node then reverses the conversion process by demultiplexing the received optical signal and converting it back to an RF signal so that it can be received by viewers.

[004] Cable television networks (CATV) have continually evolved since they were first developed as relatively simple systems that delivered video channels one way from a content provider. Early systems included transmitters that assigned a number of CATV channels to separate frequency bands, each approximately 6 MHz. Subsequent advances allowed limited feedback communication from subscribers back to the content provider via a small dedicated signal. low frequency propagated in the coaxial network. Modern CATV networks, however, provide not only a much greater quantity of content channels, but also provide data services (such as Internet access) that require much greater bandwidth to be allocated to both forward and backward trajectories. return. In the specification, drawings, and claims, the terms "front trajectory" and "downstream" may be used interchangeably to refer to a trajectory from a service center to a node, from a node to an end user. , or from a service center to an end user. In contrast, the terms "return path", "reverse path", and "upstream" may be used interchangeably to refer to a path from an end user to a node, from a node to a service desk, or from an end user to a service desk.

[005] Recent improvements in CATV architectures that provide additional improvements in content delivery include Fiber to the Premises (FTTP) architectures that replace the coaxial network between a node and a subscriber's home with a fiber optic network. Such architectures are also called Radio Frequency Through Glass (RFoG) architectures. A key benefit of RFoG is that it provides faster connection speeds and more bandwidth than current coaxial transmission paths can deliver. For example, a single copper coaxial paired conductor can carry six simultaneous phone calls, while a single fiber pair can carry more than 2.5 million phone calls at the same time. FTTP also allows consumers to bundle their communication services to receive telephone, video, audio, television, any other digital data products or services at the same time.

[006] An existing weakness of RFoG communication channels is Optical Beat Interference (OBI), which affects traditional RFoG networks. OBI occurs when two or more reverse path transmitters are switched on, and are very close in wavelength to each other. OBI limits upstream traffic, but it can also limit downstream traffic. Existing OBI mitigation efforts have focused on Optical Network Units (ONUs) at the customer premises, or the CMTS at the service desk. For example, some attempts to mitigate OBI make ONUs wavelength specific while other attempts create an RFoG-aware scheduler in the CMTS. Still other attempts have included changing UN wavelengths directly. Due to the fundamental nature of laser traffic and DOCSIS, none of the above techniques produced satisfactory results as wavelength collisions continued to occur or the cost was high. Therefore, it may be desirable in RFoG installations to further reduce or eliminate OBI.

BRIEF DESCRIPTION OF THE VARIOUS VIEWS OF THE DRAWINGS

[007] Figure 1 shows an existing RFoG architecture.

[008] Figure 2 shows an improved RFoG architecture.

[009] Figure 3 compares the capabilities of the architectures of Figures 1 and 2.

[010] Figure 4 shows an RFoG transmission path between a CMTS and a cable modem.

[011] Figure 5 shows an improved ONU that mitigates clipping.

[012] Figure 6 shows a second improved ONU that mitigates clipping.

[013] Figure 7 shows an ONU output spectrum that has a rise time of 100 ns.

[014] Figure 8 shows an ONU output spectrum that has a rise time of 1000 ns.

[015] Figure 9 shows a response time of an ONU to an RF signal.

[016] Figure 10 shows an ONU that has a laser tilt control and RF amplifier gain.

[017] Figure 11 shows the response time of an ONU with RF gain control in proportion to laser tilt control.

[018] Figure 12 shows the response time of an ONU in which the RF gain control is delayed in relation to the laser tilt control.

[019] Figure 13 shows an ONU that has a separate control for amplifier gain and laser tilt.

[020] Figure 14 shows a transmission line receiving structure.

[021] Figure 15 shows a transmission line receive connection to a tilted amplifier.

[022] Figure 16 shows a transmission line receiver with photocurrent detection on the termination side.

[023] Figure 17 shows an active combiner with multiple inputs and operation in continuous optical firing mode.

[024] Figure 18 shows an active combiner operating in continuous optical firing mode that includes amplifier tilt control.

[025] Figure 19 shows an active combiner with OBM, laser tilt, amplifier tilt and gain control.

DETAILED DESCRIPTION

[026] Figure 1 shows an exemplary existing RFoG system 10, in which a service desk 12 delivers content to a ONU 14 on a customer premises via a node 16. An RFoG topology includes an all-fiber service from the service desk 12 to a field node or optical network unit (ONU), which is typically located at or near the user's premises. At the service center 12, a downstream laser sends a broadcast signal that is optically split multiple times. The optical network unit, or ONU, covers the RF broadcast signal and passes it to the subscriber's coaxial network.

[027] The service center 12 typically includes a transmitter 18 that delivers a signal downstream to one or more 1x32 passive splitters 20 that include 32 output ports, each output port being connected to a division multiplexer splitter. wavelength (WDM) 28 that delivers content downstream over a fiber transmission segment 24 to node 16, which in turn includes another 1x32 splitter 22, where each output port of splitter 22 is connected via from another fiber segment 26 to a particular ONU 14 on a subscriber's premises.

[028] Optical network units (ONUs) in an RFoG environment terminate the fiber connection at a subscriber-side interface and convert the traffic for delivery over the home network at the customer's premises. Coaxial cable may be used to connect the ONUs of an RFoG network to one or more user devices, wherein the RFoG user devices may include cable modems, EMTAs, or signal converters, as with user devices. of an HFC network. For example, the ONU 14 may connect to signal converters, cable modems, or similar network elements via coaxial cable, and one or more of the cable modems may connect to the subscriber's internal telephone cabling and/or to personal computers or similar devices via Ethernet or Wi-Fi connections.

[029] Persons of ordinary skill in the art will appreciate that the aforementioned architecture is illustrative only. For example, the number of ports of separators 20 and 22 can be changed as desired. It should also be understood that the service center 12 may include more splitters 20, with each splitter having outputs connected to a respective node in order to serve a large number of subscribers.

[030] Along the return path from subscriber ONU 14 to service center 12, separator 22 operates as a combiner, that is, up to 32 ONUs can deliver return path signals to node 16, which combines the same for upstream transmission along the length of fiber 24. Each of the signals coming from the respective ONUs 14 is then separated from the other signals by the WDM 28 to be received by a separate receiver 30 in the service center 12. The signals coming from from respective receivers are then combined by a combiner 32 for transmission to a Cable Modem Termination Service (CMTS) at service center 12. The signals are combined in the RF domain at service center 12 by combiner 32, before to be connected to the CMTS upstream port. Due to the front-end power limit in fiber, combined signals require one front-end fiber (LI km) per group of 32 subscribers.

[031] In the forward direction, the front transmitter is supplied to a higher multiport power amplifier that distributes power. For example, in service center 12, transmitter 18 provides output to an Erbium Treated Fiber Amplifier (EDFA) 34 that distributes power internally across the 32 outputs of combiner 20, with each output being operated at a relatively high power, e.g. example, approximately 18 decibel-milliwatts (dBm). The WDM 28 typically passes 1550 nm of light from the EDFA 34 in the forward direction and directs the reverse light, typically at 1610 nm or perhaps 1310 nm in the reverse direction, to the receivers 30. The WDM 28 can be connected to a fiber of length L1 that feeds the splitter 22 at node 16. The outputs of the splitter 22 are each supplied to second fibers of length L2 which are respectively connected to ONUs 14 at the subscribers' homes. Typically, L1+L2 can be up to 25 km. The ONUs 14 convert forward transmitted light to RF signals for the coaxial network in the home. The ONUs 14 also receive RF signals from the home network and modulate these signals into a laser, which operates at 1610 nm, for example, and the laser output is sent upstream on the L2 fiber. The upstream signal is combined with other upstream signals in combiner 22 and further transmitted upstream on fiber L1. In WDM 28, upstream signals are directed to service center receivers 30.

[032] The loss budget for 32 subscribers and 25 km of fiber requires a receiver in service center 12 for each group of 32 subscribers; Given an upstream transmit power of 3 dBm, the receivers 30 and the WDM 28 can typically operate at a power between -18 and -21 dBm, which forms a good signal to challenge noise ratio, so that Limited bandwidth receivers are typically required for acceptable performance. Furthermore, the passive optical combiner 22 that combines multiple optical inputs into a single output, by definition, creates OBI between these inputs, as previously described, and will therefore create noise in the RF domain in service desk receivers 30. Furthermore Furthermore, a loss of around 24 dB must also be assumed in the frontal trajectory; for an EDFA output power of 18 dBm per port, this provides -6 dBm of power to the receivers. This is enough for acceptable performance in the ONU at 1 GHz, given the low noise, high gain receivers are used.

[033] The disclosed techniques for eliminating OBI are desirable, and the disclosed way to eliminate OBI as disclosed may allow for higher capacity upstream and downstream. Additionally, the revealed combiner and combiner capabilities may enable the coexistence of RFoG alongside traditional HFC/D3.1 systems and potential future PON systems. Eliminating OBI is critical in some systems to unlock the vast potential of fiber optics. Described in more detail herein are embodiments for an architecture that incorporates the disclosed optical combiner system.

[034] Figure 2 shows an improved system 100 for delivering CATV content to a plurality of subscribers over an RFoG network. The architecture shows a service center 110 that has a transmitter 112 and a receiver 114, each connected to a WDM splitter 116 that emits a signal to, and receives a signal from, an L1 km fiber link 118. The fiber link 118 is connected to an active splitter/combiner unit 120. The splitter/combiner unit 120 may preferably include a WDM 122 that separates forward path signals from reverse path signals. The forward path signal from the WDM 122 is provided to an EDFA 124 which outputs an amplified optical signal to a 1x32 active splitter 126 that has 32 output ports, each to respective links of the second fiber 128. At each port, the level Power output can be modest (e.g. in the range 0 to 10 dBm) but can also be high (e.g. in the range 18 dBm).

[035] In the reverse direction, the 1x32 port separator 126 operates as an active combiner 126, and includes, at each port, a WDM that directs light upstream to a detector in the port, which converts received optical signals into electrical signals, amplifies the same in the RF domain, and provides the electrical signals to a transmitter 129 which emits light at, for example, 1610 nm, 1310 nm, or some other appropriate wavelength, supplied to the WDM 122, which in turn directs the upstream light in the fiber 118. In the service center 110, the fiber 118 is connected to the WDM 116 which directs the upstream light to the receiver 114.

[036] Each of the 32 ports of the separator/combiner 126, through a respective fiber 128, outputs a respective signal to a second active separator/combiner unit 130 of the same type and configuration as the separator/combiner unit 120. length(s) of fiber 128 may vary relative to one another. The output power per separator port is low, about 0 dBm. The splitter ports are connected to ONUs 140, for example, in a Multiple Exposure Time Unit (MDU) or a neighborhood, via fiber 132 of length L3. In a basic RFoG system, the sum of L1+L2+L3 fiber lengths is up to 25 km. The system 100, however, will allow for a greater total length of fiber between the service center 110 and the ONUs 140, such as 40 km, due to the fact that the system 100 can tolerate a greater SNR loss, as further described below.

[037] Upstream signals originating from ONU 140 are individually terminated directly at the active separator/combiner unit 130; Even for ONUs operating at 0 dBm, the power reaching the detectors is about -2 dBm (the 132 fiber is a short fiber of up to a few km, and the WDM loss within the active combiner is small). This is almost 20 dB higher than in existing RFoG systems, which means that the RF levels after the detector at separator 134 is almost 40 dB higher than in existing RFoG systems. As a consequence, the receiver noise figure is not critical, and high bandwidth receivers can be used with relatively low noise performance. The received RF signal is retransmitted via transmitter 136 along the reverse path to fiber 128 and received and retransmitted by the preceding active splitter/combiner unit 120 and thereafter to service center 110. Although retransmission Repeated application leads to some incremental reduction in SNR, the SNR improvements from the active architecture provide much greater overall performance over traditional RFoG systems. Most importantly, because all inverse signals are individually terminated in separate detectors, there can be no optical beat interference (OBI) between different inverse signals. The inverse signals are not combined optically, hence OBI cannot occur.

[038] Although, in some embodiments, RF splitter/combiner units such as 120 and 130 may use an RF combiner to combine respective electrical signals from each detector at each port, this may produce unacceptable losses in transmission to amount from the UN to the service center. Therefore, the RF cutter/combiner units 120 and 130 preferably have the detectors arranged in a transmission line structure as shown in Figure 14, which will not incur such high signal loss.

[039] In the forward direction there may be multiple EDFAs, such as EDFA 124 in the separator/combiner unit 120; These EDFAs are economical single-stage devices with low power dissipation – typically 2 Watts or less. The cascade formation of the EDFAs results in an accumulation of noise due to the finite noise figures of the EDFAs. Although the active separator architecture does not require EDFAs, since an EDFA (not shown) in a high-power service center 110 can continue to be used to provide power to the ONUs 140, the use of EDFAs, such as the EDFA 124, within active separator units, it provides some advantages. For example, the complexity and power dissipation of the equipment in the service center 110 is greatly reduced, as in the fiber count emanating from the service center 110. The amount of power delivered to the ONUs 140 is readily increased to approximately 0 dBm at from -6 dBm in a traditional RFoG system. As a consequence, ONU receivers obtain 12 dB more RF level from their detectors and do not need as high a gain or as low a receiver noise contribution. Even with relaxed noise requirements on ONU receivers, the SNR impact due to EDFA noise is easily overcome due to the higher received power. Furthermore, more spectrum can be sustained in the forward direction with an acceptable SNR over current architectures, such as 4 GHz instead of 1 GHz in current RFoG, from which total data transfer rates can grow significantly without a change in operation to allow, for example, services that provide 40 Gbps download speeds and 10 Gbps upload transfer speeds.

[040] In some embodiments, the optical combiner provides upstream and downstream RFoG capability and a completely transparent and reciprocal capability for PON transmission. The optical combiner can enable complete transparency for PON installations. For example, the optical combiner can enable high-capacity, OBI-free resources by installing in HFC D3.1-capable FTTH-capable networks. Similarly, the optical combiner can be incorporated into GPON, 1G-EPON, XGPON1, 10G/1G-EPON, 10G/10G-EPON. HFC and D3.1 compatibility allows the revealed optical combiner to be installed alongside a current HFC network, and is D3.1 ready. The optical combiner can be installed in a fiber node, in a multi-exposure unit (MDU), and in single-family home installations (SFU).

[041] Embodiments for an RFoG combiner include preventing or eliminating OBI at the combiner as opposed to managing it at the edges of the network (such as using a CMTS scheduler on the service desk side of the network or specific ONUs wavelength at the subscriber end of the network). Modalities that allow for the elimination of OBI are described. The revealed optical combiner can be used to eliminate OBI, improve capacity and/or enable multiple services in RFoG, the cable version of FTTH networks.

[042] The disclosed optical combiner can be independent of ONUs, Cable Modems and CMTSs. The revealed optical combiner can be CMTS agnostic, which thus eliminates the need to create an RFoG-aware scheduler, which is both restrictive and time-consuming. The optical combiner makes a cable version of FTTH more feasible compared to PON alternatives. For example, in embodiments, the disclosed optical combiner has a reciprocal PON traversal capability of the optical combiner along with high upstream and downstream capacity, which assists in the installation of RFoG without disruption to the underlying system, or future inclusion of inability to PON functionality, such as retrofitting PON into an RFOG system.

[043] In some embodiments, the optical combiner has 32 ports, but requires only one transmit port, one receive port, on a WDM component in the service center. This way, instead of needing 32 WDMs and 32 receiving ports, the revealed optical combiner can save space and power in the service center. The combiner can be an active device that needs approximately 2 Watts of power. The optical combiner may be powered by power supplies readily available in the RFoG system, or power may be provided at the optical combiner. The power source may include a battery backup or solar/fiber power alternatives. If power is lost and the battery has also been drained, all PON reciprocal transmission is not affected. Upstream RFoG transmission, however, is stopped. In a conventional RFoG system, it was also stopped due to the fact that the preponderance of OBI would have severely weakened the system anyway if the system was a traditional RFoG system with a passive combiner. Also in the case of power loss, the ONU (Optical Network Unit) in homes would cease to function so that without any power reserve such systems will cease to function, if these are RFoG or PON systems, with or without the active combiner revealed in the present context. The service desk optical receiver 114 may only need an input power range of 0 to 3 dBm, and require 15 dB less output RF power due to the absence of the RF combiner so that with such a need for high input optical power and low output RF power, gain may be low.

[044] The disclosed optical combiner can preferably eliminate the OBI, which forms an OBI-free system. The optical combiner allows long range and large divisions, for example up to 40 km and 1024 divisions, which will be further expanded. The high upstream and downstream capacity enabled by the revealed optical combiner includes up to 10G DS/1G US, and as much as 40G DS/10G US.

[045] In embodiments, the disclosed optical combiner prevents interference with RFOG installations in the combiner rather than preventing interference with the use of measurement taken at the ONU in which previous attempts failed or proved to be cost prohibitive.

[046] Traditional RFoG architectures have a fixed power budget. This means that as the fiber length between the service center and the ONUs increases, a smaller number of splits can be used, as can be seen in Figure 3, where the lower curved line represents the existing architecture and the upper curved line represents the active architecture disclosed in this document. Conversely, the more divisions desired, the shorter the length of fiber can be installed. The disclosed active architecture, however, allows fiber lengths of up to approximately 40 km regardless of a number of splits used, which means that the disclosed active architecture allows fiber lengths of 40 km or more along with a large number of splits, e.g. example, 1024, which thereby advances FTTP topology and development.

[047] The overall cost of the active separator architecture shown in Figure 2 is similar to that of a traditional RFoG solution. The cost of the active separator EDFA gain blocks and the WDM and detector components in the active architecture is displaced by eliminating service desk gear such as receivers, high-power EDFAs, and combiners. A cost reduction of ONUs that can operate at lower output power additionally supports the active separator architecture. Additional advantages of the active splitter architecture can include a reduction in the fiber count leaving the service center, which can have a large impact on system cost, as well as an option to use 1310 nm reverse ONUs while remaining within a typical SNR loss budget, which can further reduce costs. Also, the system shown in Figure 2 exhibits increased bandwidth over what existing RFOG architectures are capable of providing, which avoids limits on service group sizes and concomitant needs for more CMTS return ports. Finally, unlike OBI mitigation techniques in existing RFoG architectures, the system shown in Figure 2 does not require cooled or temperature-controlled optics and bidirectional communication links that require additional ONU intelligence.

[048] Each of these factors provides an additional cost advantage of an active separator solution over existing RFoG architectures. The space and power required in the service center are also reduced; the active separator solution requires a transmit port, a receive port, and a WDM component. Existing RFoG architectures, on the other hand, require transmit ports, high-power EDFAs with multiple ports, 32 WDMs, 32 receive ports, and a 32-port RF combiner. Existing RFoG architectures require very low noise, high gain, output power receivers with damper methods deployed to overcome power loss and noise addition in the RF combiner. The system 100 shown in Figure 2, reciprocally, operates with input power typically in the range of 0 to 3 dBm, little gain is required, and requires 15 dB less power output due to the absence of the RF combiner beforehand. of CMTS.

[049] Preferably, the disclosed optical combiner unit deploys a transmission line approach to combine multiple optical photodetectors into a single optical receiver. This can be achieved in unidirectional and bidirectional configurations. A unidirectional system does not provide control communication signals from an active optical separator to an ONU, that is, control communication signals pass only from an ONU to an active separator. Thus, in a unidirectional system, an active optical separator simply accepts an output level from an ONU and operates at that output level. A bidirectional system passes control signals from an active optical separator to ONUs that instruct them to adjust their output power; This type of system allows precise equalization of input levels to the active optical separator from each ONU.

[050] Some active separator/combiner systems may preferably include redundancy when the active optical separators switch their return laser power (the return laser that carries the combined information from the ONUs connected to it) between a high and low power or operate this laser in CW mode. In this case, an upstream service center or active optical splitter can easily detect power loss at one input port and allow a second input port connected to another fiber route to receive the information; in the forward path, the other fiber route would also be activated in this case due to the fact that generally the forward and reverse light share the same fiber. Also, some active separator/combiner systems may include an inverse laser in the active optical separator that adjusts its power output as a function of the amount of transmitter ONUs for the active optical separator and the photocurrent received from those ONUs. Still other active separator/combiner systems may have an active optical separator inverse laser power and gain factor set to a fixed value.

[051] Preferably, the disclosed optical combiner unit is capable of configuring itself under changing circumstances. There are cases in which it is necessary for cable modems in the ONU to communicate with the CMTS even if there is no data to be transmitted. Typically, however, the ONU is turned off during periods when there is no data to be transmitted between the ONU and the CMTS, and a cable modem can take hours to receive or send data. Thus, in some embodiments, the disclosed combiner unit may be configured to remain in communication with the CMTS. Cable modems may be required to communicate back to the CMTS once every 30 seconds, or some other appropriate interval.

ONU OPERATING MODES AND LASER CUTTING PREVENTION

[052] In traditional RFoG architectures, ONUs transmit information in continuous bursts and at any point in time one or more ONUs can turn on and begin transmitting information. As required by the DOCSIS descriptive report, all ONUs are probed repeatedly at intervals of up to 5 minutes, but typically less. When an ONU is turned on, the optical power transmitted by the ONU rises from zero to the rated output power in a short time. As a consequence, the optical power received by the active separator of that ONU goes through this same transition. The rate of turnover at which the ONU can switch is restricted by the DOCSIS descriptive report, but the transition continues to be relatively abrupt, resembling a step function. As is well known from signal theory, a step function has a frequency spectrum that contains significant energy at low frequencies, with the energy falling as the frequency rises. If low frequency energy were allowed to be retransmitted unimpeded by the active splitter laser when retransmitting signals, then the signal could readily bypass the laser and cause laser clipping. To avoid such a cut-off, several approaches can be used.

[053] First, a steep high-pass filter can be deployed after the active separator detectors, which ensures that low-frequency signals induced in the photodetectors from the ONUs that turn on and off do not bypass the laser used for relay. Such a high-pass filter must be constructed so that it presents low impedance to the photodetectors for low frequencies, so that the photodetectors do not see a significant slope fluctuation when the ONUs are cycling on and off. For example, if a coupled capacitor were used as the first element in a filter that presents high impedance to the photodetectors, then an ONU turning on could result in a significant slope fluctuation of the photodetectors; such a filter should preferably not be used. In this context, a significant slope fluctuation would be a fluctuation of more than 10%. Preferably, the high pass filter is configured to limit fluctuations to levels well below this figure, for example 5% or even 2%. Also, if the relay laser is used in continuous firing mode, then the slew rate of the relay laser should be limited, preferably turned on, to limit the amount of low-frequency spectrum in the photodetectors. of previous active separator units.

[054] As noted above, ONUs typically operate in continuous firing mode and this causes the associated problems just described. Continuous firing mode operation of ONUs is required in an existing RFoG architecture due to the fact that otherwise the probability of OBI occurring would be very high and the system would generally not function. With the active separator architecture, however, OBI cannot occur and the signal to noise margin is much higher than with RFoG. Because of this, a second approach to reducing clipping is to operate ONUs in a continuous “on” state with the active architecture described previously. For 32 ONUs delivering signals into an active separator, shot noise and laser noise accumulate, but the signal to noise budget is so high that the resulting SNR performance continues to be much better over RFoG systems. existing. As a consequence, the active separator architecture allows the operation of all connected ONUs at the same time, as the active separator architecture eliminates OBI.

[055] A third option to alleviate laser clipping is to allow the ONUs to operate in continuous firing mode, but to sense an amount of power outside the ONU and attenuate the signal from the ONU so as to prevent clipping. Referring to Figure 4, using a traditional RFoG system 200, the CMTS 210 can maintain the RF level at a constant feedback input port. The return signal is generated by a cable modem 220, supplied to an ONU 230 that includes an inverse optical transmitter, and relayed over an optical network 240 to a receiver 250 co-located with the CMTS that converts the optical signal back to a signal signal. RF and provides this to the CMTS 210. It should be understood that the optical network 240 may contain active and passive elements. It should also be understood that communication between the cable modem 220 and the CMTS 210 is bidirectional, that is, there are both “forward” and “reverse” trajectory signals.

[056] The communication path shown in Figure 4 can be used to adjust the output level of the cable modem 220. In case the loss from the ONU 230 to the receiver 250 is high, or the loss from the receiver 250 to the CMTS 210 is high, so the CMTS 210 will adjust the output level of the cable modem 220 to a high level in order to obtain an input level defined in the CMTS or a level within a predefined range in the CMTS. In traditional RFoG systems, there is considerable headroom in the input level that the ONU can handle, to allow for this adjustment. However, it is still possible for the cable modem 220 to exceed the ONU 230, particularly as the amount of spectrum used by the cable modem increases to support future heavy data loads. When the ONU 230 is over-expressed, then the modulated RF signal in the ONU 230 laser becomes so high that the inverse laser in the ONU 230 is triggered at the cutout, i.e., the laser output power varies so low that the laser is turned off. This causes severe signal distortions and creates a wide spectrum of frequencies that interfere with communication across the entire spectrum.

[057] The optical network typically combines signals from multiple ONUs, each ONU is typically communicating in another band of the frequency spectrum. The communication of all these ONUs is affected by the wide spectrum induced by the distortions even if only one ONU is cut out. Preferably, this problem is resolved in such a way that the other ONUs are not affected, the cropped ONU is placed in a state where it can continue to communicate, and the CMTS produces a warning that a ONU does not operate in such a way. ideal.

[058] A variation on the third option just described is to operate the ONUs in continuous firing mode in which the ONU switches between a low power state (e.g. -6 dBm) and a high power state (e.g. 0 dBm). This means that the ONU laser never turns off completely, i.e. the laser always operates above its laser threshold, and can always be monitored by the active separator. The reduction in output power when it does not transmit RF signals reduces triggering and laser noise accumulated in the active separator so that the signal to noise impact is minimized.

[059] In circumstances in which the optical combiner unit rotates to a low power state rather than a completely off state, the photodiode current and a max/min can be traced to the photodiode current across all ports of the combiner, and thus a microcontroller can be used in the optical combiner to continuously track the max and min in a specified time interval. For example, if for ten minutes the max photodiode current is 0, then the optical combiner determines that the cable modem is also not connected, has a faulty optical link, or is otherwise defective. Optionally, the active optical combiner can signal the absence of photocurrent to a service center. The optical combiner also has the ability to configure itself whether or not the optical combiner can determine whether the incoming light is burst, as in normal RFoG operation, or CW (continuous wave) as with a node inverse transmitter. The optical combiner has the ability to use CMTS upstream signaling imposed by CMTS on modems to analyze which ports are working, which ports are silent, which input ports are connected to ONUs, and which input ports are connected to optical combiner inverse transmitters, where the optical combiner ports may have a different output power profile than the ONUs in that the power may be CW or may be fluctuating between a low and high power state or may carry information embedded in signage indicating the presence of an additional optical combiner between the ONU and the optical combiner.

[060] For active cascade separators, the feedback lasers in active cascade separators can be operated similarly in conventional continuous firing mode in which the laser turns off between continuous firings, in CW mode, or in a trigger mode continuous switch that switches between a high and low power state. It should also be understood that CW operation of inverse lasers and/or ONUs, or continuous firing mode operation with a low and a high level further facilitates the determination of optical input levels at the input ports upstream of the separators. active. It should also be understood that although the devices and methods disclosed in the present application that prevent or otherwise reduce clipping by a laser operating in continuous firing mode have been described in the context of a ONU, the devices and methods used to preventing clipping by a laser in an ONU are equally applicable to preventing clipping by a laser in an active separator as previously disclosed.

[061] Figure 5 shows a system that mitigates laser clipping that may result from otherwise continuous firing mode communications from an ONU. Specifically, an ONU 300 may include an RF rms detector 310, a microcontroller 320, and an algorithm for adjusting an attenuator 330 in the ONU as a result of power detected in the RF rms detector 310. The reverse trajectory from the ONU 300 can be operated in continuous shooting mode; When an RF signal is presented to input 340, then the ONU laser 350 is turned on by tilt circuit 360. This can be accomplished either by an additional RF detector (not shown in Figure) in the input circuit that directly turns on the tilt circuit (dashed arrow) or by the RF detector 310 and the microcontroller 320 that turn on the tilt and set the tilt level. When a continuous trigger occurs, the RF detector 310 measures a power level and provides this to the microcontroller 320. The microcontroller is also aware of the operating current of the laser 350 as defined by the slope circuit 360. In this way, the microcontroller 320 can compute whether the RF level signal is large enough to induce inverse laser clipping. If no clipping occurs, no further action needs to be taken and the ONU 300 can retain a nominal RF attenuation value. If at this time the ONU is not at a nominal RF attenuation value, the procedure is more complicated, this will be discussed later in the specification.

[062] If clipping occurs, the microcontroller 320 stores the event. If a specified number of clipping events have been counted within a specified time interval, then the microcontroller 320 determines that the ONU 300 has significant performance degradation due to the clipping, and also significantly disables other ONUs in the system. In this case, the microcontroller 320 computes how much the RF attenuation needs to be increased to eliminate clipping using RF power measurements that were previously recorded. The microcontroller 320 then increases the RF attenuation to a new value so that the laser 350 is modulated more strongly than normal (with more modulation index than the nominal value), but remains below the cutoff. The microcontroller 320 can also optionally increase the laser slope setting to provide more tolerance for laser modulation.

[063] Due to the fact that the signal attenuation from the ONU 300 has been increased, the RF level as seen by the CMTS at the end of the link drops. The CMTS will then attempt to instruct the cable modem to increase the output level to restore the desired input level to the CMTS. This can result in one of two scenarios. First, the cable modem may not have the capacity to increase the output level further and the CMTS will list the cable modem as a problem unit that does not have the capacity to obtain the desired input level to the CMTS. This does not mean that the CMTS can no longer receive signals from the cable modem, as the CMTS has a wide input range to accept signals. Thereafter, the reverse trajectory will generally continue to function while it would have been severely weakened if the clipping problem had not been resolved. Second, the cable modem may have more tolerance, in which case the CMTS will instruct it to increase its output level and restore the CMTS input level to the desired value. As a consequence, the reverse laser will be driven into the cutout again and the ONU microcontroller will further increase the RF attenuation. This cycle will continue until the cable modem has reached its maximum output capacity, and then the system is back to the first scenario.

[064] The system shown in Figure 5 provides protection against clipping by ONUs, and also makes the CMTS aware of problem modems or ONUs. As noted previously, the root cause of the problem was that the loss from the ONU to the CMTS was too great, due to, for example, a poor fiber connection in the optical network from the ONU to the receiver. This issue is flagged and will eventually be resolved. When the problem is resolved, however, the CMTS input level increases beyond the preferred CMTS input level, and the CMTS then directs the cable modem to reduce the output level. If the ONU is not at the nominal attenuation value and you notice that the actual modulation index is o below the nominal level, then this can be recognized as different from the previous “new value” for ONUs that were over-expressed, this was deliberately set above the nominal modulation index. This implies that the problem in the system has been resolved and the microcontroller can reduce the attenuation to the nominal value, gradually or in one step. In this way, this technique automatically recovers from the state in which it protects the ONU from clipping with increasing attenuation to the nominal attenuation once the system has been repaired.

[065] As previously indicated, an ONU takes time to turn on after a continuous shot has been detected. For example, the RFoG descriptive report indicates that the turn-on time of an ONU should be between 100ns and 1000ns (i.e. 1µs). A turn-on time that is too fast undesirably creates a very high low-frequency noise, which decreases as the frequency increases. Unfortunately, because this noise extends to about 50MHz or beyond, most currently upstream installable signals propagate within the frequency range that is affected by noise due to abrupt turn-on time. Exacerbating signal degradation is the fact that noise is spikey, where the continuous shot of instantaneous noise can be much louder than what is typically seen on a spectrum analyzer with moderate video bandwidth.

[066] Figure 6 generally illustrates an upstream architecture of ONU 400 in which an RF detector 410 detects whether an RF signal is present at its input 420. If a signal is detected, the RF detector 410 passes the signal to an amplifier 450 and also signals a laser tilt control module 430 to turn on at time t0 a laser 440, which has an on time 460. The amplifier 450 amplifies the RF signal that is passed through the RF detection circuit 410 The amplified signal drives the laser 440. The laser output is propagated from the ONU on a fiber 470. For simplicity, the downstream ONU receiving architecture is not shown in Figure 6. The laser turn-on time 460 has a profound effect on the spectrum produced by the ligation event.

[067] Figures 7 and 8 show the estimated spectra for a rise time of 100 ns and 1 µß, respectively, for a typical signal at 40 MHz. For a short rise time, the noise due to the ONU connection is of the same order of magnitude as the intended signal. With a slower laser ligation, this effect can be mitigated.

[068] If there is only one ONU turned on at any given point in time, the effect of low frequency noise due to the ONU turning on is negligible, due to the fact that the DOCSIS load is inserted after the laser has completely turned on. However, when there are multiple ONUs that can be switched on at any given time, then the noise is often non-negligible. If there was a first ONU turned on and a second ONU turned on while the first one was transmitting data, then the high noise peaks described above are present over a wide range of the frequency spectrum of the upstream signal while the first ONU was on. transmitting data. Depending on the relative RF levels of the signals and the magnitude of the noise peaks, the signal may experience pre- or even post-forward error correction (FEC) type errors when measured in the CMTS, for example. The potential for deliberate noise becomes more and more pronounced as the numbers of ONUs that can be linked increases as is likely to happen as architectures migrate to the DOCSIS 3.1 standard. Although this problem has always existed, it only becomes apparent, as a basis for residual error, when the OBI and its induced errors are eliminated.

[069] An additional weakness is caused by applying the RF signal before the laser is completely turned on and has stabilized. Specifically, a weakness may occur, for example, if the laser on time is shorter than the DOCSIS Preamble that may be applied before the laser has reached steady state. Typically, the DOCSIS Preamble is sent as a QPSK signal and can often be 6 to 10 dB louder than the next regular RF signal, depending on signal conditions. In such an instance, the laser will be overexpressed while remaining in a low power state and experiencing very large clipping events that can cause spikes in the noise throughout the entire RF spectrum of the upstream signal, and thus obscure other signs that may exist at the same time. As previously indicated, although this effect has always occurred, it only becomes observable with the elimination of OBI, if its OBI-induced errors are present.

[070] Figure 9 shows an inclination, close to which a laser is modulated with a sine wave signal. During the time that the laser tilt is insufficient, the output signal is clipped. For a slower laser ligation, the clipping duration is increased. Although it may be desirable to reduce low-frequency RF spikes that occur along the upstream frequency spectrum that has a slower turn-on time, the increase in clipping described above may counteract the benefit of the slow turn-on time. Innovative techniques are revealed that allow for slow ligation time while avoiding clipping artifacts.

[071] Referring to Figure 10, an innovative ONU upstream architecture 500 includes an RF detector 510 that detects whether an RF signal is present at its input 520. If a signal is detected, the RF detector 510 passes the signal to an amplifier 550 and also signals a laser tilt control module 530 to turn on at time t0 a laser 540, which has an on time 560. The laser tilt control module 530 preferably modulates the tilt of the laser 540 to achieve a complete turn-on of the laser 540 for a turn-on time 560 that is preferably as slow as possible, e.g., the slowest turn-on time allowed by the RFoG standard or, in some embodiments, even longer . In some embodiments, the on time of the laser 540 may be up to 500ns, 1µß or more. This can greatly reduce low-frequency noise. The turn-on time for the laser can be linear, as shown in Figure 10, or it can deploy a transition along any other desired curve, such as a polynomial curve, an exponential curve, a logarithmic curve, or any other desired response.

[072] The amplifier 550 amplifies the RF signal that is passed through the RF detection circuit 510. The amplified signal drives the laser 540. Preferably, when amplifying the RF signal from the RF detector 510, the laser tilt control 530 includes a circuit that modulates amplifier gain to be proportional to laser tilt. This effectively sets the gain of the amplifier 550 as proportional to the laser link 560, and thereby reduces or even prevents overtriggering and clipping by the laser 540. The laser output is then propagated from the ONU in a fiber 570.

[073] Figure 11 shows the output of the laser 540 when using the system in Figure 10. As seen in this Figure, when using an RF gain factor proportional to the laser inclination, clipping no longer occurs. However, variation in the RF level during laser turn-on can potentially cause a problem in the burst receiver that can expect an approximately constant RF level during laser turn-on. To mitigate this, in some embodiments, the amplifier slope may be modulated to delay the RF signal to the laser, relative to the laser on time 540, and may also apply a constant time faster than the optical power on time. This modality is illustrated in Figure 12.

[074] Figure 13 shows an implementation of an ONU that includes a delay in the RF signal to the laser, in relation to the laser turn-on time, and also applies a constant time faster than the optical power turn-on time. Specifically, an innovative ONU upstream architecture 600 includes an RF detector 610 that detects whether an RF signal is present at its input 620. If a signal is detected, the RF detector 610 passes the signal to an amplifier 650 and also signals a laser/amplifier tilt control module 630 to turn on at time t0 a laser 640, which has an on time 660. The laser/amplifier tilt control module 630 preferably modulates the tilt of the laser 640 to achieve a complete turn-on of the laser 640 for a turn-on time 660 that is preferably as slow as possible, e.g., the slowest turn-on time allowed by the RFoG standard or, in some embodiments, even longer. In some embodiments, the on time of the laser 640 can be up to 500ns, 1µß or more. This can greatly reduce low-frequency noise. The turn-on time for the laser can be linear, as shown in Figure 13, or it can deploy a transition along any other desired curve, such as a polynomial curve, an exponential curve, a logarithmic curve, or any other desired response.

[075] The amplifier 650 amplifies the RF signal that is passed through the RF detection circuit 610. The amplified signal drives the laser 640. Preferably, when amplifying the RF signal from the RF detector 610, the laser/amplifier tilt control 630 includes a circuit that modulates the amplifier gain to be proportional to the laser tilt, but with a delay 680 relative to the time at which the laser 640 starts to turn on. Preferably, the amplifier gain rise time is faster than the laser link rise time. In some embodiments, the laser/amplifier tilt control module 630 simply turns on the RF gain, i.e., rise time as short as the amplifier allows. The laser output is then propagated from the ONU on a 670 fiber.

[076] This ONU shown in Figure 13 effectively sets the gain of the amplifier 650 as proportional to the laser linkage 660, and thereby reduces, or even prevents, overtriggering and clipping by the laser 640, while at the same time mitigating problems caused by a receiver expecting a nearly constant RF level during the time the laser is turned on. The ability to simultaneously reduce laser on time and provide RF gain to the laser in proportion to the laser on time but delayed relative to the laser on time is a feature that has great potential. in all applications and, without loss of generality, these techniques can be used for any analogous application such as DOCISIS 3.0 or 3.1.

[077] One of (or both) the architectures shown in Figures 10 and 13 can be used together with the architecture shown in Figure 2 in order to further improve the speed and stability of HFC systems. These can be used in addition to the long-term clipping reduction discussed in the previous disclosure to reduce both long-term and short-term clipping effects on the system. CONTINUOUS SHOOTING DETECTION

[078] As previously indicated, upstream transmissions typically operate in continuous firing (BM) mode, in which the ONUs energize a transmitter, for example a laser, only during intervals of time when information must be transmitted to the along the upstream trajectory. A system in continuous shooting mode generally provides a lower noise environment and thus allows better SNR, and in the case of the transmitter being an optical device, the use of continuous shooting mode tends to reduce Optical Beat Interference (OBI ). Thus, in some preferred embodiments of the previously disclosed optical combiner system disclosed in this specification, in which the OBI must be suppressed, such optical combiners are preferably operated in continuous firing mode.

[079] Also as previously indicated, RFoG architectures that use continuous firing mode detect the RF level at the ONU, turning on the ONU laser when an RF signal is detected and turning off the laser when the RF signal is not detected. gift. This procedure is referred to as “RF detection”. In an optical combiner, the optical light inputs coming from the ONUs are all detected and the detector outputs are collected. If RF detection is used with an optical combiner, an RF comparator would be applied to the output of the combined RF output. If the RF level output from the combined RF detectors was higher than the applied comparator, then the optical laser in the optical combiner would be disabled.

[080] However, such detection may experience difficulties due to the fact that the input RF level may be very small. For example, a small slice of a D3.1 signal can be produced by any single ONU, hence the modulation index of the ONU would be low, which results in a low RF level in the optical combiner. Also, the optical input power to the optical combiner from a ONU may be low; With an optical input range reaching up to 12 dB, the RF level upon detection can vary by 24 dB. As a result, the RF level from a photodiode may continue to be so low that the RF level that should be detected would be lower than the comparator, even if the RF level were high relative to the Optical Modulation Index of the photodiode. ONU laser that generated the RF signal. In ONU embodiments, the RF level may be turned on after the optical output is turned on, or while the optical output is being turned on, so that detection of an RF level in the revealed optical combiner is delayed. Furthermore, detection can also be slow due to the fact that it depends on the comparator circuit.

[081] An alternative to using continuous shot detection in the cascaded optical combiner units disclosed in the present application would be to maintain upstream light transmission at all times, regardless of whether signals are supplied to the optical combiner or not, i.e. , an “always on optical combiner”. While this may ensure that the optical combiner transparently retransmits upstream information, it would result in constant light input to all ports in an upstream optical combiner device or multiport receiver. Full light entering the ports in this way can lead to a sum of noise triggered from all ports, which degrades the SNR performance of the total system. Therefore, in preferred embodiments, the optical combiner unit transmits light upstream only when an RF signal has been received and is to be sent.

[082] Disclosed herein is an innovative continuous shot detection method that is fast, simple, stable and robust, which thus allows for multiple new architectures. Specifically, as widely stated, the disclosed optical combiner system can monitor the optical current of each photodiode as well as the sum current of all photodiodes. If any of the photodiodes register a photo current, or alternatively a current above a certain minimum value, the relay laser is automatically switched on. Photodiode current generation is instantaneous and beneficially it is a DC value that is easier to compare. As interconnection network speeds increase over time, such optical detection circuits will become more useful.

[083] Such an Optical Continuous Firing Mode (OBM) detector promotes reliability and may have the following advantages: (1) in the case of multiple optical combiners connected in series in a linear fashion as disclosed in the present application, the substantial reduction in triggered noise of addition is achieved relative to an “always on” solution; (2) in the case of DOCSIS 3.1 transmission, individual signal transmissions with very low RF levels per ONU can be detected and retransmitted; and (3) in the case of varying optical input levels due to different optical lengths between the ONUs and the disclosed active optical combiner, or varying optical lengths between linearly series-connected multiples of such active optical combiners, reliable operation in Continuous shooting mode can continue to be achieved.

[084] Furthermore, the disclosed innovative burst detection also allows detection of input light immediately at the start of a burst burst at the optical combiner input. Conversely, when there is no light at the input, or alternatively, no light for a certain period of time, the auxiliary RF amplifiers in the disclosed active optical combiner can be turned off, thereby reducing the power dissipation of the active optical combiner. revealed. When light appears at the input of the revealed active optical combiner, the amplifiers can be turned on again within the allowed time; for example, in an RFoG system up to a microsecond is allowed to establish an optical link from the moment the RF input is detected and the system begins to power up. Due to the fact that RF amplifiers take a finite time to turn on and establish amplification; Early detection of a continuous trip is important to provide sufficient time to establish normal operation. Such power cycle termination can reduce power dissipation by as much as tenfold, thereby dramatically improving critical infrastructure metrics. In this way, for example, in the case of a power output, the optical combiner can conserve the required power not only by using continuous optical trigger operation, but also circuitry for RF continuous trigger operation, and extend the life of a battery, if available.

[085] The implementation of an optical power detection circuit with the capacity to cover a wide range of optical input power, in an architecture that has multiple detectors is not trivial. Given the large number of detectors present, combined with a wide range of optical input power, the amount and range of photocurrent that needs to be reliably detected is considerable. Simply measuring the voltage drop across a resistor in the detector's slope network is difficult; At low input power on a single detector, a small voltage drop can be reliably detected only if the value of a resistor, across which is a voltage drop equal to the photodetector slope, is relatively high. However, increasing the value of such a resistor is not desirable due to the fact that it leads to an increased voltage drop when high detector currents are present in multiple detectors; detector tilt would become a strong function of the optical light present in the detectors. In some embodiments, the slope of the detector is kept constant due to the fact that the responsiveness of the detector depends on the slope of the detector; Therefore, a variation in the detector's inclination can lead to a variation in the system gain. Even a resistance value as low as a typical transmission line impedance, such as 75 Ohms, can be problematic when a large number of detectors are active, and, for example, 100 mA of detector current flows in the multiple detector system, which takes to an excessive drop in detector inclination.

[086] A method is disclosed for detecting optical light over a wide range of input power while retaining a constant slope in detectors present in the transmission line receiver. In order to achieve this, a combination of an RF amplifier and a transimpedance amplifier are used with the multiple detector structures. In some embodiments, the transimpedance amplifier is connected to a high-pass structure in front of the RF amplifier so that, for low frequencies, the transimpedance amplifier has a very low impedance connection (lower than the transmission line impedance ) to the detector inclination.

[087] Referring to Figure 14, which shows an example of a transmission line receiving structure 700, a photodetector can be accurately shaped up to reasonably high frequencies (up to 1 GHz) by a capacitance in parallel with a source of current for reasonable input power levels (>1 uW). Thus, in this Figure, each of the circuit elements 710 would be a model of a photodetector. Conventional receiver models use a transimpedance amplifier or detector match with as high an impedance as possible, such as 300 Ohms, in order to convert the current source signal into an RF signal with the best possible noise performance. These approaches are limited by detector capacitance so that an increase in the number of detectors by simply combining detectors or detector area leads to a loss of detector performance due to an increase in combined detector capacitance and is therefore unreasonable to expect A large number of detectors (for example 32) are expected to work well with a single RF amplifier. This implies that multiple amplifiers are needed to receive a large number of fibers.

[088] As an alternative, multiple detectors can be supplied to an RF combiner before being amplified. An RF combiner requires each detector to be individually terminated with an RF impedance that is typically less than 100 Ohms, which will consume half the detector current and, due to combined signals coming from multiple detectors, the RF combiner will introduce additional power loss. at least 10*log(N) dB, where N is the number of combined detectors. This loss becomes excessive for 8 detectors or more. Additionally, other losses are caused by practical implementations of RF combiners that require costly transformers to carry out. Transformers also cause bandwidth limitations and other aforementioned losses, and are difficult to deploy for high impedances (such as more than 100 Ohms).

[089] In the disclosed transmission line receiver, use is made of the insight that an inversely tilted photodetector behaves as a current source in parallel to a capacitor with a low loss at RF frequencies. This transmission line receiver will not induce 10*log(N) RF combiner loss, does not require transformers, offers high bandwidth, and is capable of providing an output signal that represents a delayed sum of a large number of detectors. A transmission line with impedance Z can be modeled by a slope network of inductors and capacitors with L/C=Z2, which works well for frequencies below the resonance frequency of L and C. Practical detector capacitance values are in the order of 0.6 pF, so a 75 Ohm transmission line would need L=3.4 nH. The resonant frequency is well above 1 GHz so that, up to 1 GHz, a transmission line with an arbitrary number of detectors compensated by 3.4 nH inductors would simulate a 75 Ohm transmission line. The quality of the parasitic capacitance of inversely inclined detectors is such that they can be considered low-loss capacitors at RF frequencies. The 3.4 nH can also be distributed near the detectors as 2x1.7 nH, which leads to a model as shown in Figure 14.

[090] As indicated above, each source/capacitor current combination 710 represents a detector. Figure 14 shows a number of these sections in series separated by the respective transmission line 720 (100 psec or on the order of 1 cm on the board) which has 75 Ohms of impedance. The detectors are matched with 1.7 nH 730 inductors. A 75 Ohm 740 resistor terminates the transmission line input. Transmission line output 750 feeds a low noise 75 Ohm RF amplifier (not shown). It should be understood that although Figure 14 shows six detectors, there is no limit to the number of detectors that can be combined by concatenating these sections, and up to the LC resonance frequency there is negligible impact on the achievable bandwidth for a large number of detectors. In practice, the 1.7 nH inductors can be implemented in the PCB layout as narrower line sections, and a balanced transmission line with 100 Ohm or 150 Ohm differential impedance can be used to slightly improve the noise figure.

[091] As shown in Figure 14, each source/capacitor current combination 710 represents a photodetector, where the current source is the current detected in the detector; and the capacitor represents the parasitic capacitance of the detector. Multiple detectors are connected to sections of the transmission line (such as T2) and corresponding inductors (such as L1 and L2). The matching inductors are chosen so that the parasitic capacitance of the photodetectors is matched to the transmission line impedance (typically 75 Ohms). In this way, multiple detectors can be connected and concatenated to a transmission line, so that detector currents are supplied to the transmission line and these detector currents are equally divided to propagate to the output 750 and the termination resistor 740 on the other. end of the transmission line structure. Each detector current generally passes through transmission line sections, corresponding inductors, and detector terminals before reaching one end of the transmission line. In this way, signals from adjacent detectors affect the signal voltages present at each detector terminal and can therefore affect the detector current itself, which causes cross-modulation of detector signals. However, because an inversely tilted detector can be modeled as a good current source, such cross-modulation does not occur. Each half of the detector current is thus presented at the transmission line output as a signal with a delay proportional to the distance from the detector to the transmission line output. The distance determines the delay from an electrical signal at the detector terminal to the transmission line output and includes delay due to matching inductors and photodetector capacitance. The signal at the transmission line output is, therefore, proportional to a sum of halves of the delayed detector current, regardless of the number of detectors in the transmission line structure. The signal at the output of the transmission line can thus be said to represent a sum of the delayed detector currents.

[092] The bandwidth of the transmission line structure is limited only by the inductive matching of the photodiode capacitance and can be very large, exceeding 1 GHz. Output 750 is connected to an RF amplifier matched to the transmission line impedance , which amplifies the signals emitted from the transmission line structure. It should be noted that the use of a transimpedance amplifier that is not matched to the transmission line structure would cause too much reflection of the output signals back to the transmission line structure; a transimpedance amplifier is not a preferred means of amplifying the output of a transmission line receiver.

[093] Typically, photodetectors need to be tilted, for example, with 5 V. In order to decouple the tilt voltage from the amplifier, a decoupling capacitor can typically be used. Tilt can then be provided via an inductor in a tilted beacon arrangement as shown in Figure 15, for example. The signal from transmission line 760 is supplied to an amplifier (not shown) through a capacitor (770) that passes high frequency signals, and slope from a voltage source 775 is supplied to the transmission line through of a 780 inductor that passes low frequency signals. The terminating resistor 740 at the other end of the transmission line is thus capacitance-decoupled to allow for a DC slope. The current through the voltage source 775 can be measured to determine photocurrent; The voltage source 775 can be interpreted as a transimpedance amplifier that provides a constant voltage and an output proportional to the current supplied. However, in implementations, the 780 inductor needs to be chosen with a value large enough so that it does not affect the low frequency response of the amplifier. As a consequence, there may be a delay in the response of the current in inductor 780 to a change in the detecting photocurrent, and this tends to cause a delay in photocurrent detection.

[094] Figure 16 shows an implementation 800 that uses both ends in the transmission line receiving structure to alleviate such delay. Resistor R1 in Figure 16 is the terminating resistor 740 shown in Figure 14, and inductor L1 is inductor 780 in Figure 15. Voltage source 810 provides slope for both terminating resistor 740 and inductor 780. current in resistor 740 responds instantly to a photocurrent so that rapid detection of photocurrent is skill. Inductor 780 can sustain large photocurrents without a significant voltage drop so that large photocurrents can be sustained without a significant drop in bias to the photodetectors. A capacitance 815 may be placed adjacent to the voltage source 810; For an ideal voltage source, it may not carry any current due to the fact that the voltage is constant. However, at RF frequencies, it may be difficult to achieve a perfect voltage source, so the capacitor 815 provides a low impedance to ground so that RF currents in the termination resistor 740 do not cause voltage modulation at the source. voltage 810.

[095] In order to realize an efficient detection circuit for the current in the voltage source 810, the voltage source 810 is preferably implemented as a transimpedance amplifier. A transimpedance amplifier is a basic electronic circuit that holds a node between two current paths at a constant voltage and has an output that changes its output voltage in proportion to the current supplied at that node. Thus, externally, the transimpedance amplifier looks like a voltage source for that node, but there is an additional output that represents the current supplied. This output can then be used to drive a decision circuit to decide whether a photocurrent flows or not. Due to the fact that the transimpedance amplifier is made with a practical transistor circuit, it does not have infinite bandwidth, which means that it does not have the ability to maintain constant node voltage for very high frequencies and therefore For this reason, capacitor 815 can be added in some embodiments.

[096] It should be understood that, in some embodiments, the LC slope network before the amplifier (capacitor 770 and inductor 780) can be replaced with more complex circuits, or even with diplex filters - as long as the network provides a path of high frequency with low loss from the transmission line detector to the amplifier, and a low loss (low impedance) path at low frequency from the voltage source (transimpedance amplifier) to the slope of the transmission line detector. It should also be noted that the transimpedance amplifier can be deployed so that the output voltage first changes linearly as a function of photocurrent, but then saturates at a photocurrent that is sufficiently high.

[097] In other implementations, a photocurrent detection circuit may be applied to each individual photodetector; Optionally, one electrode of a photodetector (e.g., cathode) may be connected to an RF circuit and the other electrode (e.g., anode) may be connected to an optical power detection circuit. This increases complexity as one detection circuit is required per detector. Also, some embodiments may optionally use a transimpedance amplifier per detector.

[098] With a detection circuit in continuous optical firing mode, for example, of the type described above, the slope of a laser or the slope or gain of an amplifier can be controlled. Figure 17 shows a multi-detector receiver 820 that produces an output 825 that signals that power has been detected from any of the multiple inputs 830. This detection may be based on a detection method as described in the previous section or on circuitry. with multiple detectors monitoring individual detectors 835. When the optical input has been detected at time t0, then the laser tilt is turned on with a controlled rise time t_on_1 and the active combiner can retransmit signals present at the inputs.

[099] Continuous optical trigger mode detection can be additionally used to control amplifier tilt as shown in Figure 18; when optical power is detected at t0 the amplifiers are immediately turned on. The laser is turned on more slowly so that the amplifiers are established for as long as the optical power is on. Optionally, this scheme can be expanded by a third control signal 850 that controls the amplifier gain, AS SHOWN IN FIGURE 19.

OPTICAL MODULATION INDEX AND SELF-CALIBRATION

[100] For deployments that allow operation of all inputs upstream of the active separator at the same time, the total amount of photocurrent in the detectors after the upstream inputs can be high. The impedance of the sloped circuit and, as discussed, the previously mentioned filtering means of the detector output path must be low.

[101] In an existing RFoG system, the CMTS controls the output level of communications from cable modems to ONUs that transmit RF signals to a service center so that a desired input level to the CMTS is obtained. This implies that the output level from a receiver preceding the CMTS is adjusted to a known level. If this receiver is of a type that has a known amount of gain such that an output level corresponds to a known index of optical modulation, then this implies that the index of optical modulation of channels supplied to the CMTS is known given the signal of RF level to which the CMTS adjusts the channel. This requires a calibrated receiver that adjusts its gain as a function of optical input level (2 dB increase in gain for every dB reduction in optical input level) so that this fixed relationship between RF output level and optical input is maintained. The modulation index at the receiver is the modulation index of the upstream laser in the active separator connected to that receiver; thus, the CMTS implicitly controls the modulation index of this active separator output.

[102] The gain of the active separator should preferably be set so that an output modulation index from that active separator has a known relationship to an input modulation index at one or more of the photodetectors receiving upstream signals from of active separators or ONUs further downstream. This requires knowledge of the photocurrents in these photodetectors and, ideally, the active separator can monitor the photocurrent of each upstream link using one detector per upstream link as in a transmission line detector, for example. Since some systems can operate in continuous shooting mode, these photocurrents are not always available. However, in a DOCSIS system all ONUs are probed repeatedly for an acknowledgment signal with an interval of up to five minutes. This implies that active upstream tabs relay the information, and all active tabs in such a system have each of the upstream inputs activated at least once every five minutes. The active separator can thus record continuous trigger levels and assemble a map of optical input levels to input ports. Using this information, the active separator can set an internal gain level so that the upstream modulation rate is maximized, but will not be clipped as long as the input signals to the active separator are not clipped. Although the length of fiber from the service center to the first active splitter is generally long, these lengths of fiber between active splitters and these lengths of fiber from active splitters to ONUs are generally short, and have low enough loss that the values of optical input power to the different upstream input ports are close together, and the ideal configuration is similar for all ports. As a consequence, the optimal gain setting on the active separator is almost the same for all input ports and the compromise in SNR of assuming the worst case inverse laser modulation index from a signal on any of the input ports is entrance is small.

[103] As noted previously, one embodiment may use high and low output optical power configuration for the inverse laser, rather than switching the laser between a high output power for continuous shot transmission and an off state in between. . Not only does this modality provide active separators with continuous information about link loss to the ONU, it also improves laser operation. When a laser is turned on, the transient leads to a brief transition where laser distortion is high and the RF input signals can be clipped. If a laser is maintained at a low power level rather than being in the off state before being switched to a higher power level, then this transient is almost absent and distortions and clipping are reduced. In case, the laser is maintained at a high output power continuously, these transients and distortions are absent. The active separator architecture allows operating the ONUs in any of these three modes and an optimal one can be selected for system operation.

[104] Although the input power levels upstream to the detectors in an active separator are typically similar, in some examples they may differ due to differences in connector loss or fiber loss. Preferably, all optical inputs would have the same level or have the same RF level after the detector for an equivalent channel load. Since the active separator can monitor the power level at each detector and map these optical input levels, it can compute adjustments to the optical input power level or modulation index of these inputs that would be necessary to equalize the RF levels. after the detectors at each entrance. The active separator can communicate these preferred settings for output power level or gain to the inverse downstream transmitters that are connected to the inputs. Communication signals can be modulated into a laser injected into downstream signals or into pump laser currents in EDFAs that amplify downstream signals. The modulation can be selected to be small enough, and in such a frequency range, that the communication signals do not interfere with the downstream content.

[105] Preferably, not only active separators receive and interpret these communication signals, but also downstream ONU units receive and interpret the signals. This would essentially allow perfect alignment of the optical transmission level and RF gain of all units in an active separator system. Given the presence of an upstream laser, and the ability of all components in an active separator system to receive an upstream signal, all components in an active separator system have the capability for upstream communication with the addition of a modulation single tone or other scheme. In this way, bidirectional communication is allowed, and the active separators and the service center can communicate with each other, the system self-exploration, and gain optimal configuration and optical levels.

[106] One objective of the active separator architecture is to provide accurate RF levels to the CMTS that represent an optical modulation index. Doing so is non-trivial, and requires a specific self-calibration procedure (described later) that is expected to result in accurate modulation correlation ratio to active splitter service center receiver output RF levels. The receiver is either a CMTS plug-in or be connected directly to the CMTS with no unknown RF loss contributions therebetween (in case a cap is required for other services in relation to the CMTS, the cap can be integrated into the receiver to avoid external RF losses). As a consequence, the modulation index of the active separator relay units is precisely defined.

[107] In case bi-directional communication is not available, then the output power level of the ONU cannot be adjusted by the active splitter and the modulation index of the ONU will continue to have some uncertainty as the optical loss between the ONU and active separator/receiver may vary; a +/- 1dB loss variation from the ONU to the active separator would result in a +/2 dB tolerance in the RF level, so a dynamic window will have to accommodate at least this variation and tolerance for other tolerances and accuracy of CMTS configuration. This must be readily available for bandwidths up to 200 MHz so that even without the active separator controlling the output power of the ONU, acceptable system performance can be achieved. With the previously mentioned bidirectional control, additional system tolerance can be achieved.

[108] When the 1200 MHz return bandwidth is used, so that the ONUs receive 200 MHz spectrum widths, the ONUs can all be operated at a few dB below their cutoff point, that is, only the enough to cover the uncertainty in the loss of the ONU to the active separator to avoid clipping the ONUs. This optimizes the performance of the critical link from the ONU to the active separator, so that 0 dBm ONUs are sufficient. In this type of operation, an arbitrary choice can be made for a number of ONUs that operate with such a 200 MHz band, for example, up to six ONUs. This in turn would cause clipping in the active separator transmitter; Thus, for 1200 MHz operation, the gain of the active separator receivers after the ONUs can be reduced by 8 dB so that when six ONUs transmit 200 MHz of signal bandwidth, the inverse active separator transmitter is operated just below the cutout. This method of operation maximizes SNR and eliminates uncertainty - the impact of variance on the link between the ONU and active separator is minimized, and active separator links are operated with a precise modulation index as with wide RF return systems. lower band. The required dynamic window is reduced to tolerances in CMTS level setting and active separator output level calibration, which allows operation at an optimal modulation index.

[109] Analysis of the SNR achievable using the system just described, for 1200 MHz operation with a maximum load of 200 MHz per ONU, results in a 5 dB improvement in the SNR achievable at 1200 MHz. This results in about 20% more throughput capacity on the system. With 1200 MHz of bandwidth, the total upstream data rate can be as high as 10 Gbs.

[110] In the event that the system is initially configured so that the active separator units expect a return spectrum of 1200 MHz (instead of, for example, 200 MHz) with a maximum of 200 MHz per ONU, then a penalty of about 7 dB occurs in terms of peak NPR performance. Therefore, the preferred mode of operation can be switched between normal operation, where a single ONU can occupy the entire spectrum, and high-bandwidth operation where a single ONU can be assigned to a limited amount of spectrum at any time. and inverse active splitter transmitters support the entire spectrum at once.

[111] The proposed architecture has multiple relay links that are preferably operated at the best possible modulation index, assuming perfect alignment of the NPR (Noise Power Ratio) curves of these links. As noted previously, retransmission alignment on the active separator return links is critical to achieving the best possible performance (each dB of misalignment directly results in a reduction in available SNR) and therefore technical calibration is required to define and maintain correct alignment of transmitter gain factors.

[112] In order to provide such calibration, the active separator feedback transmitter gain will be precisely defined such that for a given detector current of the active separator receiver diodes, the modulation index of the transmitter is equal to the input from the modulation index to the detector. This only requires knowledge of the detector current; the actual optical input power to the detector and the detector's responsiveness are irrelevant. In order to achieve this, means are implanted in each detector to measure the detector current so that an appropriate gain can be set for the feedback transmitter.

[113] The gain can be set individually for each detector, but since multiple detectors can be receiving signals at the same time, this would require a controllable attenuator for each detector (32 detectors are in a typical active separator unit). Preferably, a single attenuator is used for all detectors. This is achieved through the use of variable output transmitters on active separator units that communicate with an upstream active separator or variable output transmitters on ONUs that communicate with an upstream active separator. Structured below is a method for defining the output level of each of the inverse transmitters so that each transmitter provides the same photocurrent in the detector to which it is coupled. During normal operation, the active splitter receiver monitors the detector currents during continuous firing to enable a warning to be issued in the event of an optical link degrading or being lost.

[114] For a 1310 nm reverse link from the active separator to an upstream active separator, the reverse laser power needs to be controlled typically either 3 to 10 dBm or 6 to 10 dBm, depending on the separator receiver model active. For a 1610 nm reverse link, these figures are typically 3 to 7 dBm or 6 to 7 dBm, respectively. These controls ensure that the received power at the end of a 25 km link, with some WDM loss, is at least 0 dBm. It should be understood that the numbers given are examples. The active separator can transmit information in the forward direction through pump modulation of the EDFA or injection of a signal into the forward path. The latter is more costly; the former results in a lower data rate, as only minimal pump fluctuation can be allowed without affecting the forward trajectory. A low data rate is sufficient, and can be read by a simple receiver - for example, a remote controller receiver operating in the kHz range coupled to a low-cost processor. It must be understood that the downstream transmission function is only required on the upstream active separator units unless the ONUs are being controlled as well. In the Figures shown, they would be one of the 33 active separator units in the system.

[115] In a self-calibration run, the upstream active separator unit transmits a downstream command to active separator units to initiate self-calibration. Subsequently, the downstream units turn their transmitters on and off randomly at full power with a low duty cycle, so that in almost all cases most downstream units are on. The upstream active separator reports information downstream as to which port is on, and what detector current was obtained from that unit. Downstream units record that information in non-volatile memory; Since they can correlate the messages to their own activities, this provides information to the downstream unit as to which port is on and what power has been supplied to that port. After all ports have been on at least once, or an output time has occurred (for example, if one or more ports are not connected), the upstream active separator unit determines which downstream active separator produces the lowest current detector. The upstream active separator then computes how the upstream powers of each of the downstream units should be set so that all detector currents are the same and fall within a specific range. This range may correspond, for example, to input power of 0 to 3 dBm (or 6 dBm) at the detectors. It should be understood that this can be achieved by setting up a photodetector current, and does not require the measurement of an exact optical input power.

[116] In general, the upstream active separator unit will set this power to the best (or maximum) value that can be obtained to optimize the links. The active separator units will then all have a known output power, and their internal gain will be set accordingly to have a modulation index calibrated to a given input power and modulation index. All links in an upstream active separator can behave identically. The upstream active separator unit can then take the downstream units out of calibration mode.

[117] In the event that an additional port is lit on an upstream active separator receiving port, then the self-calibration algorithm can proceed without interruption of service of already connected active separator units. This is achieved by activating self-calibration on the receiver of the downstream active separator that was activated just now by requesting calibration mode only for units with unknown port number (i.e. only the new unit). Its output will be turned on and the upstream active separator unit will then assign a port number to the new, and previously unused, port and set a power to the new unit, and take it out of calibration mode.

[118] During normal operation, the upstream active separator unit continues to monitor the receiving currents for the upstream input links. If there is a significant deviation, this may continue to issue a downstream command in non-calibration mode to readjust power, and may also signal upstream plant problems.

[119] Active separator units operated in the disclosed manner may also assemble a map of connected active separator units. Also, a map of an upstream power can be created from the connected ONUs and statistics about individual ONU operation and link loss can be collected, for example, to locate chatting ONUs or weak ONU connections.

[120] The service center transmitter may also send a command to downstream active separator units to initiate calibration or change an operating mode (e.g., from 200 MHz to 1200 MHz of optimized operation). Any other type of bidirectional EMS monitoring system can be designed for active separator units that can receive and transmit low data rate traffic. It must be understood that this does not need complex and costly HFC EMS systems; secondary optical power fluctuations either by pump power variation or low-level signal injection in the downstream signal path, or inverse laser power variation in the upstream path, are sufficient to detect binary or kHz range (such as remote control chips) of modulated data patterns. It must also be understood that the most costly option - injection of an optical signal downstream - is only relevant in the service center, or in the upstream path typically only relevant at 1 of 33 active separator locations. Another important consideration is that the CMTS must set modem levels correctly. In regular feedback or RFoG systems, there is considerable uncertainty at the system levels due to applied RF components or combiner networks. In the active separator system, however, there are no RF components on the link, the service group is aggregated in the optical domain, and only a low gain, low performance, and low output level receiver is required, which is coupled directly to the CMTS return port. In some embodiments, it may be desirable to produce a dedicated active separator receiver with an output level precisely calibrated as a function of the input modulation index. Such a receiver has no need for a wide input range; from -3 to +3 (or from 0 to +6) dBm is enough. The high input level implies that the gain may be low. The absence of RF combination after the receiver also means that the output level can be low. Therefore, such a receiver must be obtainable in a high-density, low-power form factor. With such a receiver, little, if any, RF cabling may be required in the service center, and the CMTS can precisely set the inverse levels to obtain the correct optical modulation index. In some cases, there may be a need to connect other equipment in relation to the CMTS to the reverse trajectory. The receiver may use an auxiliary output to provide this functionality, rather than the main output with external RF separators. This eliminates any level uncertainty due to RF components between the receiver and CMTS.

MODALITIES

[121] Some embodiments of the aforementioned disclosure may encompass multiple cascaded active separators that are configured to work with ONUs based primarily on optical input levels without requiring bidirectional communication. Other embodiments may encompass multiple cascaded active separators that are configured to work with ONUs using bidirectional communication.

[122] Some embodiments of the aforementioned disclosure may include an active separator with multiple optical inputs, each of which provides an optical input to one or more detectors that, together, output a combined signal to a high-pass filter that has a low impedance to detectors and rejects all signals below an RF frequency range and passes all signals above an RF frequency range before presenting the combined signal to an amplifier and a relay laser.

[123] Some embodiments of the aforementioned disclosure may include an active separator with multiple optical inputs, each of which provides an optical input to one or more detectors that together output a combined signal, wherein the active separator has a tilted circuit with a sufficiently low impedance at low frequency so that all detectors can be illuminated at the same time without a significant drop in slope to the detectors.

[124] Some embodiments of the foregoing disclosure may include an active separator with an inverse laser in which the inverse laser turns on when a photocurrent in the input detectors of the active separator is above a threshold, and in which the rate of rotation when the activation of the laser is limited so that it does not create a transient that has a spectrum that interferes with the upstream spectrum being transmitted.

[125] Some embodiments of the foregoing disclosure may include an RFoG active separator architecture in which the inverse lasers of the active separator(s) and/or ONUs connected to the active separator(s) ) are operated with a continuous output. Some embodiments of the foregoing disclosure may include an RFoG active separator architecture in which lasers inverting the active separator(s) and/or ONUs connected to the active separator(s) are operated between a high and low power mode so that the output power is high during upstream transmission bursts and is otherwise low at the output. Some embodiments of the foregoing disclosure may include an RFoG active separator architecture in which inverse lasers of the active separator(s) and/or ONUs connected to the active separator(s) may be selectively defined. to one of a continuous mode and a continuous shooting mode.

[126] Some embodiments of the foregoing disclosure may include an RFoG ONU that switches between a high and low output power state in which the output power is high during burst transmission of information and in which a low output power state is output is above the laser threshold.

[127] Some embodiments of the aforementioned disclosure may include an RFoG system that measures detector currents at all inputs, assemble a table of detector currents during high and low input power (or not) to the optical inputs, and compute based on that table , a gain mode value that a modulation index of the reverse transmission laser has a known relationship to a modulation index at the optical inputs to the active separator, so that the reverse transmission laser has an ideal modulation index, but the clipping is prevented, even for the port with the highest optical input. In some embodiments of the aforementioned disclosure, the ideal modulation index of the inverse transmitter is nominally the same as that for the optical inputs.

[128] Some embodiments of the foregoing disclosure may include an RFoG ONU with an RF signal detector that detects continuous bursts of input signals and activates a laser in a high power mode when a continuous burst is detected and otherwise active. the laser in a low power mode, such as zero power. An electrical attenuator may precede the laser driver and may attenuate an RF input signal, so that in the low output power state the laser cannot be clipped by an RF input signal. The RF attenuation before the laser can be reduced as the laser power increases from the low power state, so that the RF attenuation is removed quickly to have minimal impact on continuous firing, but during the transition the laser continues not to be cropped.

[129] Some embodiments of the foregoing disclosure may include an RFoG ONU with an RF signal detector that detects continuous bursts of input signals and includes an electrical attenuator that precedes the laser driver to attenuate the RF input signal, so so that when no nominal input is present, noise funneling by the ONU of weak noise signals at the ONU is prevented and RF attenuation is quickly removed when a continuous shot is detected as having minimal impact on the continuous shot.

[130] Some embodiments of the foregoing disclosure may include an RFoG ONU that may receive a downstream signal that instructs it to adjust the output power level, RF gain, or both. In some embodiments, such ONU may be assigned port numbers and status monitoring requests. In some embodiments, such ONU may transmit upstream information such as status, serial number, etc.

[131] Some embodiments of the aforementioned disclosure may include an active separator that may transmit a signal downstream with requests for downstream units to adjust the level of optical power, gain, or to request status information. Some embodiments may include an active separator that can receive such signals downstream. Some embodiments may include an active separator that can transmit and/or receive such signals in the upstream direction, as well.

[132] Some embodiments of the foregoing disclosure may include an ONU with an RF detector, an attenuator, a bias circuit, and a microcontroller in which the microcontroller estimates laser clipping based on measured RF power levels and tracks at which fraction of time the laser is clipped and increases attenuation if this fraction exceeds a threshold. The microcontroller can also adjust the laser tilt to prevent clipping. In some embodiments, the microcontroller brings the attenuation to a nominal value when the RF power to the laser is at or below a nominal value. In some embodiments, changes in attenuation made by the microcontroller occur in discrete steps in time and magnitude.

[133] In some embodiments of the aforementioned disclosure, the microcontroller may set the attenuation to a level high enough to prevent clipping, but less than necessary to obtain a nominal modulation index.

[134] Some embodiments of the foregoing disclosure may include a bidirectional fiber RF architecture with more than one relay link in the reverse direction, wherein signals detected from preceding links are combined on each relay link.

[135] Some embodiments of the foregoing disclosure may include a calibrated receiver in a service center that provides a specific RF output level for a modulation index input, with a gain control such that, for different optical input levels, the RF output level for a given modulation index is kept constant. In some embodiments, a receiver may include two outputs, at least one connected to a CMTS without any combination and separation of RF networks.

[136] Some embodiments of the foregoing disclosure may include an active separator with at least two gain settings, a gain setting optimized for ONUs that can transmit the full inverse spectrum that the system can sustain, and a setting optimized for ONUs that can transmit an amount of spectrum less than the total spectrum the system can sustain, where the active splitter combines inputs from multiple ONUs and can transmit the total spectrum the system can sustain.

[137] Some embodiments of the foregoing disclosure may include an active separator that has adjustable inverse transmit power and adjustable gain so that, for a given received upstream signal modulation rate, the active separator maintains a constant optical modulation rate independently of the output optical power. In some embodiments, the transmitted optical modulation index is the same as the received optical modulation index. In some embodiments, the transmitted optical modulation index is a predetermined fraction of the received optical modulation index, and the separator allows an option to vary this fraction.

[138] Some embodiments of the aforementioned disclosure may include an active separator that can receive and decode front-end communication signals, for example an input monitoring diode for an EDFA, or another monitoring diode.

[139] Some embodiments of the aforementioned disclosure may include an active separator that can transmit forward communication signals, with, for example, a forward laser, or by modulating the pump current of an EDFA.

[140] Some embodiments of the aforementioned disclosure may include an active separator that can receive and decode upstream communication signals, for example, by monitoring upstream detector currents. Some embodiments of the foregoing disclosure may include an active separator that can transmit upstream communication signals, for example, by inversely modulating the laser.

[141] Some embodiments of the aforementioned disclosure may include a system with at least two active separators in which a first active separator instructs a second active separator to adjust its inverse transmit power level. Some embodiments may use an algorithm to equalize and optimize the inverse transmission level of all downstream active separators connected to an upstream active separator. In some embodiments, the algorithm runs automatically at startup so that the downstream active separators (and optionally the ONUs) obtain an address and optionally report in the upstream direction the serial number and status of the separator (or the ONU). ). In some embodiments, subsequent activation of ports on the separator leads to automatic calibration of new ports without interrupting service of existing ports, and with continuous monitoring of port integrity.

[142] Some embodiments of the foregoing disclosure may include an active separator capable of upstream communication, and capable of receiving and decoding upstream communications from another separator.

[143] In some embodiments, an active separator may establish a map of the system in which it is included, and may report system status and topology information to a service center and may issue alarms if necessary. The map may include serial quantities of active separators, and may include serial quantities of connected ONUs. Some embodiments may create a system map automatically, and (i) may monitor ONU link input levels for active splitters; (ii) may detect vibration or otherwise defective ONUs and, optionally, instruct the active separator to turn off detectors of defective or talking ONUs; and/or (iii) can monitor the status of the active separator that builds the map. In some embodiments, the monitoring function is used to automatically trigger route redundancy by monitoring upstream traffic on a link to determine whether the link is intact, and if it is concluded that the link is faulty, switch the downstream traffic. to an alternative upstream link. In some embodiments, the upstream active separators monitor downstream active separators by communicating with the downstream active separators.

[144] Some embodiments of the aforementioned disclosure may include a service center that instructs active downstream separators to initiate a self-calibration procedure.

[145] Some embodiments include a combiner that can monitor each of the upstream input ports and thereby detect a loss of a link to that port. The loss of an upstream link implies that the associated downstream link has been lost. Detection of a link can be used to initiate switching to a redundant fiber link, preferably following a different fiber route.

[146] The terms and expressions that were used in the aforementioned specification are used therein as terms of description and not of limitation, and there is no intention, in the use of such terms and expressions, to exclude equivalents of the features shown and described or portions of the same, it being recognized that the scope of the claimed subject matter is defined and limited only by the claims that follow.

Claims (26)

1. Active separator, CHARACTERIZED by the fact that it comprises: a unit that operates as a separator in a forward direction and an active combiner in a reverse direction, the combiner having a plurality of inputs each of which receives an optical signal to amount from a respective ONU and combines them to create a combined electrical signal at an output; a transmitter that receives the combined electrical signal and converts the combined signal into a reverse path optical signal, the transmitter adjusting at least one of output and power gain to maintain a constant ratio between the optical modulation index at the output and the optical modulation index at an input, wherein the active separator has a first gain setting optimized for Optical Network Units (ONUs) transmitting a first inverse spectrum and a second gain setting optimized for ONUs transmitting an inverse spectrum limited to a amount smaller than the first inverse spectrum, wherein the active separator is configured to transmit the first inverse spectrum; and a multiplexer that multiplexes the reverse path optical signal with a forward path optical signal. 2. Active separator, according to claim 1, CHARACTERIZED by the fact that the constant ratio of the optical modulation index at the output is maintained independently of the respective optical power received at the plurality of inputs. 3. Active separator according to claim 2, CHARACTERIZED by the fact that the optical modulation index at the output is set to the optical modulation index of the plurality of inputs. 4. Active separator according to claim 2, CHARACTERIZED by the fact that the optical modulation index at the output is set to an adjustable fraction of the optical modulation index of the plurality of inputs. 5. System comprising a service center, a plurality of ONUs and at least one active separator, the active separator being CHARACTERIZED by the fact that it comprises: a unit that operates as a separator in a forward direction and an active combiner in a reverse direction, wherein the combiner has a plurality of inputs each of which receives an upstream optical signal from a respective ONU and combines them to create a combined electrical signal at an output; a transmitter that receives the combined electrical signal and converts it into a reverse path optical signal at a calibrated power level; a multiplexer that multiplexes the reverse path optical signal with a forward path optical signal; wherein the active separator, at least one of: (i) calibrates at least one of its own reverse transmit optical power level and RF signal gain to that of at least one other device, and in response to a downstream signal ; (ii) calibrates at least one of its own inverse transmit optical power level and RF signal gain to that of at least one other device, and in response to an upstream signal; and (iii) calibrates at least one of a reverse transmit optical power level and RF signal gain of another device using a downstream signal sent from the active separator to the other device. 6. System, according to claim 5, CHARACTERIZED by the fact that the active separator has the ability to decode received forward trajectory communication signals. 7. System according to claim 5, CHARACTERIZED by the fact that the active separator has the ability to receive a forward trajectory communication signal, modulate the received forward trajectory communication signal, and output the modulated forward trajectory signal to another device. 8. System according to claim 5, CHARACTERIZED by the fact that the active separator has the ability to instruct a second active separator to adjust its inverse transmit power level. 9. System, according to claim 5, CHARACTERIZED by the fact that the active separator constructs a system map that comprises topology information, wherein the topology information includes at least one of: (i) serial quantities of others active separators operatively connected to the active separator; (ii) series quantities of ONUs operationally connected to the active separator. 10. System, according to claim 9, CHARACTERIZED by the fact that said active separator has the ability to use the map to report system status information and said topology information to a service center. 11. System, according to claim 10, CHARACTERIZED by the fact that said system status information includes the status of said active separator. 12. System, according to claim 10, CHARACTERIZED by the fact that an upstream active separator communicates with and monitors an active downstream separator. 13. System, according to claim 9, CHARACTERIZED by the fact that said active separator has the ability to detect at least one of an input level to an active separator from a ONU, and ONUs in conversation. 14. System, according to claim 13, CHARACTERIZED by the fact that it has the ability to disable detectors of at least one of the defective ONUs and ONUs in conversation. 15. System according to claim 9, CHARACTERIZED by the fact that the active separator has the ability to automatically trigger route redundancy by monitoring upstream transmission over a first link and switching upstream traffic to a second link if the first link is not operational. 16. System, according to claim 5, CHARACTERIZED by the fact that the service center has the ability to instruct an active separator to initiate a self-calibration procedure. 17. System according to claim 5, CHARACTERIZED by the fact that the inverse transmission level of all other active separators connected and downstream of said active separator are equalized to an optimized level. 18. System, according to claim 17, CHARACTERIZED by the fact that said inverse transmission level of any active downstream separator is equalized to said optimized level automatically upon startup of said downstream separator. 19. System according to claim 6, CHARACTERIZED by the fact that the active separator has the ability to receive and decode an upstream signal, and communicate with an upstream device using the decoded upstream signal. 20. System, according to claim 5, CHARACTERIZED by the fact that calibration occurs automatically. 21. System, according to claim 5, CHARACTERIZED by the fact that the at least one Optical Network Unit (ONU) has the capacity to adjust at least one of output power level and gain, in response to an instruction received by the said UN in a downstream signal. 22. System, according to claim 5, CHARACTERIZED by the fact that at least one ONU can receive an assigned port number. 23. System, according to claim 5, CHARACTERIZED by the fact that at least one UN can respond to a situation monitoring request. 24. System, according to claim 5, CHARACTERIZED by the fact that the at least one ONU can send its serial number and/or status information to an upstream device. 25. System, according to claim 5, CHARACTERIZED by the fact that the service center includes a calibrated receiver that provides a constant RF output level based on a modulation index of an input signal, adjusting a gain in RF output level as a function of optical input level. 26. System according to claim 25, CHARACTERIZED by the fact that the calibrated receiver has two RF outputs, at least one connected directly to a CMTS that does not have any RF combination network and does not have any separation network of RF.